SAP: Business data cloud for AI

With the Business Data Cloud, SAP is making a new attempt to consolidate and standardize data from SAP's business applications as well as from third-party providers. In cooperation with data lake specialist Databricks, this includes the analysis of large volumes of unstructured data. This should enable companies to exploit the full potential of data and provide them with a robust basis for building and operating reliable agent-based AI applications.

The Business Data Cloud is similar to SAP's data management platform Datasphere, but is marketed as a SaaS product. The functions, including the Databricks modules, are provisioned via SAP's PaaS environment BTP (Business Technology Platform). This means that services no longer have to be laboriously plugged together and tenants no longer have to be managed. Another new feature is the Knowledge Graph, which maps the relationships between data and processes in a business context.

Understanding the data included

With the Business Data Cloud, SAP wants to get involved in the data product economy. Fully managed SAP data products are provided for all types of business processes. These provide access to value-creating information without costly data extraction processes. The original business context and semantics are retained. For example, if a CFO wants to assess the impact of rising inflation on profitability, the data cloud integrates external real-time data such as the consumer price index with internal financial data such as general ledger accounts, cost centers or VAT types to provide a comprehensive financial snapshot.

SAP also wants to offer new functions with the Business Data Cloud. The so-called Insight Apps link the data products with AI models to enable further-reaching analyses and planning for day-to-day business operations. SAP's GenAI co-pilot environment Joule also benefits from the offerings on the new data cloud when it comes to orchestrating AI agents across various business functions.

A detailed blog post from SAP and Databricks explains the details.

(vbr)